0 Replies Latest reply on Sep 10, 2008 10:59 AM by Jack Terranova

    RichFaces 3.1.6 does not work with STATE_SAVING_METHOD=SERVE

    Jack Terranova Newbie

      I have seen a couple of posts related to https://jira.jboss.org/jira/browse/RF-3542 where the fix in 3.1.6 seems to break server side state saving.

      In my case, as in others that I have read, the application(s) will come up, screen will render but links/navigation do not work.

      The problem started with an issue in 3.1.4/3.1.5 (prob earlier too) where AjaxStateManager was grabbing an old view and restoring after a nav case that used a redirect. A new view should have been created. The fix (as I have seen in the 3.1.6 source) comments out code that does this old view fetch and seems to completely break server side state saving.

      In a RF 3.1.4 deployment, I created my own state manager (extending AjaxStateManager) and basically recreated the bug fix by commenting out the old view fetch. This has the same affect - breaking my application.

      I have seen posts like these over at Oracle forums where a user had the same problem with an OC4j env. One poster claims that it had something to do with styles used to render components - no idea if that actually solved anything.

      My case is a RAD env (WAS 6.1), JSF 1.1 using facelets.

      Yes, I know I can workaround this by not using redirects, but removing those is not my call.

      I can't stay on 3.1.4 since redirects end up restoring old views (ones that use keepAlive and thus basically turn a request bean into a session) and 3.1.6 obvisouly just isn't working. That I have not seen a bevy of posts related to 3.1.6 and server side state saving tells me this issue is limited and is probably environmental.

      Can someone please please help piece this puzzle together?